Plan listings don't react on page and per_page

[antoninrykalsky]

Listing all plans returns 100 items. API does not respond to page or per_page query parameters.

Found while developing an ansible collection,
equinix/ansible-collection-equinix#146

[t0mk]

example check:

curl -H "X-Auth-Token: $METAL_AUTH_TOKEN"  "https://api.equinix.com/metal/v1/plans?per_page=90" | jq ".plans | length"

[t0mk]

Page doesn't change:

Page doesn't change
curl -H "X-Auth-Token: $METAL_AUTH_TOKEN"  "https://api.equinix.com/metal/v1/plans?page=2" | md5sum -
curl -H "X-Auth-Token: $METAL_AUTH_TOKEN"  "https://api.equinix.com/metal/v1/plans?page=1" | md5sum -

[t0mk]

type probably works:

curl -s -H "X-Auth-Token: $METAL_AUTH_TOKEN"  "https://api.equinix.com/metal/v1/plans?type=standard" | jq ".plans | length"
15

[ctreatma]

I only get 63 plans, which indicates that I'm getting all plans back; it's possible that the same is happening in your case, and there really are exactly 100 plans available to you.

I also see that the API is not responding to pagination params, but on closer inspection, according to the spec, that endpoint doesn't support pagination parameters and the response schema doesn't include pagination metadata.

For categories, it appears that the spec is incorrect; the parameter name should be categories[]:

curl -H "X-Auth-Token: $METAL_AUTH_TOKEN"  "https://api.equinix.com/metal/v1/plans?categories[]=storage" | jq ".plans | length"

[t0mk]

Oh I see what's going on there for me. I get duplicate entries in the plans listing, two entries are the same, they differ only in slug:

Duplicate 0:

{"id": "b2c4d757-d9b9-58cd-8380-393e0f4a16cb", "slug": "w3tintel.8358.512.6080",...}

Duplicate 1:

{"id": "b2c4d757-d9b9-58cd-8380-393e0f4a16cb", "slug": "w3tintel.8358.512.6080.x86",...}

There's 68 unique entries for me. I put python script that verifies this in next comment.

[t0mk]

plans.json

#!/usr/bin/env python3

import json

with open("plans.json") as f:
    plans = json.load(f)['plans']


print("Number of plans:", len(plans))

ids = [p['id'] for p in plans]
print("Number of unique plans IDs:", len(set(ids)))

last_id = ids[-1]

dupes = [p for p in plans if p['id'] == last_id]

print("Number of plans with ID %s: %d" % (last_id, len(dupes)))

for i, d in enumerate(dupes):
    print("Duplicate %d:" % i)
    print(d)

print()

for k in dupes[0]:
    v0 = dupes[0][k]
    v1 = dupes[1][k]
    if v0 != v1:
        print("Same-ID plans differ in \"%s\": \"%s\" != \"%s\"" % (k, v0, v1))

print()


d = {}

for i in set(ids):
    ps = [p for p in plans if p['id'] == i]
    n = len(ps)
    d[ps[0]['slug']] = n
print("Number of plans in listing per slug:")
print("===================================")
for k, v in d.items():
    print(k, v)

[displague]

Considering how unconventional this syntax is for the EM API, this may be an upstream defect. Another reason to question these results is that only one plan takes advantage of categories to report compute or legacy_gen.

$ emget '/plans?categories[]=compute' | jq -e '.plans[] .slug'
"c3.small.x86"